SURE-tools 2.1.4__tar.gz → 2.1.6__tar.gz

{sure_tools-2.1.4 → sure_tools-2.1.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SURE-tools
-Version: 2.1.4
+Version: 2.1.6
 Summary: Succinct Representation of Single Cells
 Home-page: https://github.com/ZengFLab/SURE
 Author: Feng Zeng

{sure_tools-2.1.4 → sure_tools-2.1.6}/SURE/PerturbFlow.py

@@ -55,44 +55,6 @@ def set_random_seed(seed):
     pyro.set_rng_seed(seed)
 
 class PerturbFlow(nn.Module):
-    """SUccinct REpresentation of single-omics cells
-
-    Parameters
-    ----------
-    inpute_size
-        Number of features (e.g., genes, peaks, proteins, etc.) per cell.
-    codebook_size
-        Number of metacells.
-    cell_factor_size
-        Number of cell-level factors. 
-    z_dim
-        Dimensionality of latent states and metacells. 
-    hidden_layers
-        A list give the numbers of neurons for each hidden layer.
-    loss_func
-        The likelihood model for single-cell data generation. 
-        
-        One of the following: 
-        * ``'negbinomial'`` -  negative binomial distribution (default)
-        * ``'poisson'`` - poisson distribution
-        * ``'multinomial'`` - multinomial distribution
-    z_dist
-        The distribution model for latent states. 
-        
-        One of the following:
-        * ``'normal'`` - normal distribution
-        * ``'laplacian'`` - Laplacian distribution
-        * ``'studentt'`` - Student-t distribution. 
-    use_cuda
-        A boolean option for switching on cuda device. 
-
-    Examples
-    --------
-    >>>
-    >>>
-    >>>
-
-    """
     def __init__(self,
                  input_size: int,
                  codebook_size: int = 200,
@@ -447,7 +409,6 @@ class PerturbFlow(nn.Module):
                 zns = pyro.sample('zn', dist.Gumbel(zn_loc, zn_scale).to_event(1))
 
             if self.cell_factor_size>0:
-                #zus = self.decoder_undesired([zns,us])
                 zus = None
                 for i in np.arange(self.cell_factor_size):
                     if i==0:
@@ -643,7 +604,6 @@ class PerturbFlow(nn.Module):
                     zns = pyro.sample('zn', dist.Gumbel(zn_loc, zn_scale).to_event(1), obs=embeds)
 
             if self.cell_factor_size>0:
-                #zus = self.decoder_undesired([zns,us])
                 zus = None
                 for i in np.arange(self.cell_factor_size):
                     if i==0:
@@ -710,7 +670,7 @@ class PerturbFlow(nn.Module):
                              xs, 
                              batch_size: int = 1024):
         """
-        Return cells' latent representations
+        Return cells' basis latent representations
 
         Parameters
         ----------
@@ -731,7 +691,7 @@ class PerturbFlow(nn.Module):
         Z = []
         with tqdm(total=len(dataloader), desc='', unit='batch') as pbar:
             for X_batch, _ in dataloader:
-                zns = self._get_cell_embedding(X_batch)
+                zns = self._get_basis_embedding(X_batch)
                 Z.append(tensor_to_numpy(zns))
                 pbar.update(1)
 
@@ -858,7 +818,7 @@ class PerturbFlow(nn.Module):
                              delta_zs, 
                              batch_size: int = 1024):
         """
-        Return cells' changes in the latent space induced by specific perturbation of a factor
+        Return cells' changes in the feature space induced by specific perturbation of a factor
 
         """
         delta_zs = convert_to_tensor(delta_zs, device=self.get_device())
@@ -1083,24 +1043,18 @@ def parse_args():
         help="the file for the record of cell-level factors",
     )
     parser.add_argument(
-        "-delta",
-        "--delta",
-        default=0.0,
-        type=float,
-        help="penalty weight for zero inflation loss",
-    )
-    parser.add_argument(
-        "-64",
-        "--float64",
-        action="store_true",
-        help="use double float precision",
+        "-bs",
+        "--batch-size",
+        default=1000,
+        type=int,
+        help="number of cells to be considered in a batch",
     )
     parser.add_argument(
-        "--z-dist",
-        default='normal',
-        type=str,
-        choices=['normal','laplacian','studentt','cauchy'],
-        help="distribution model for latent representation",
+        "-lr",
+        "--learning-rate",
+        default=0.0001,
+        type=float,
+        help="learning rate for Adam optimizer",
     )
     parser.add_argument(
         "-cs",
@@ -1109,6 +1063,13 @@ def parse_args():
         type=int,
         help="size of vector quantization codebook",
     )
+    parser.add_argument(
+        "--z-dist",
+        default='gumbel',
+        type=str,
+        choices=['normal','laplacian','studentt','gumbel','cauchy'],
+        help="distribution model for latent representation",
+    )
     parser.add_argument(
         "-zd",
         "--z-dim",
@@ -1116,6 +1077,27 @@ def parse_args():
         type=int,
         help="size of the tensor representing the latent variable z variable",
     )
+    parser.add_argument(
+        "-likeli",
+        "--likelihood",
+        default='negbinomial',
+        type=str,
+        choices=['negbinomial', 'multinomial', 'poisson', 'bernoulli'],
+        help="specify the distribution likelihood function",
+    )
+    parser.add_argument(
+        "-zi",
+        "--zeroinflate",
+        action="store_true",
+        help="use zero-inflated estimation",
+    )
+    parser.add_argument(
+        "-id",
+        "--inverse-dispersion",
+        default=10.0,
+        type=float,
+        help="inverse dispersion prior for negative binomial",
+    )
     parser.add_argument(
         "-hl",
         "--hidden-layers",
@@ -1150,18 +1132,10 @@ def parse_args():
         help="post functions for activation layers, could be none or dropout, default is 'none'",
     )
     parser.add_argument(
-        "-id",
-        "--inverse-dispersion",
-        default=10.0,
-        type=float,
-        help="inverse dispersion prior for negative binomial",
-    )
-    parser.add_argument(
-        "-lr",
-        "--learning-rate",
-        default=0.0001,
-        type=float,
-        help="learning rate for Adam optimizer",
+        "-64",
+        "--float64",
+        action="store_true",
+        help="use double float precision",
     )
     parser.add_argument(
         "-dr",
@@ -1182,50 +1156,7 @@ def parse_args():
         default=0.95,
         type=float,
         help="beta-1 parameter for Adam optimizer",
-    )
-    parser.add_argument(
-        "-bs",
-        "--batch-size",
-        default=1000,
-        type=int,
-        help="number of cells to be considered in a batch",
-    )
-    parser.add_argument(
-        "-gp",
-        "--gate-prior",
-        default=0.6,
-        type=float,
-        help="gate prior for zero-inflated model",
-    )
-    parser.add_argument(
-        "-likeli",
-        "--likelihood",
-        default='negbinomial',
-        type=str,
-        choices=['negbinomial', 'multinomial', 'poisson', 'gaussian','lognormal'],
-        help="specify the distribution likelihood function",
-    )
-    parser.add_argument(
-        "-dirichlet",
-        "--use-dirichlet",
-        action="store_true",
-        help="use Dirichlet distribution over gene frequency",
-    )
-    parser.add_argument(
-        "-mass",
-        "--dirichlet-mass",
-        default=1,
-        type=float,
-        help="mass param for dirichlet model",
-    )
-    parser.add_argument(
-        "-zi",
-        "--zero-inflation",
-        default='exact',
-        type=str,
-        choices=['none','exact','inexact'],
-        help="use zero-inflated estimation",
-    )
+    )  
     parser.add_argument(
         "--seed",
         default=None,
@@ -1266,29 +1197,23 @@ def main():
     input_size = xs.shape[1]
     cell_factor_size = 0 if us is None else us.shape[1]
 
-    latent_dist = args.z_dist
-
     ###########################################
     perturbflow = PerturbFlow(
         input_size=input_size,
         cell_factor_size=cell_factor_size,
         inverse_dispersion=args.inverse_dispersion,
-        latent_dim=args.latent_dim,
+        z_dim=args.z_dim,
         hidden_layers=args.hidden_layers,
         hidden_layer_activation=args.hidden_layer_activation,
         use_cuda=args.cuda,
         config_enum=args.enum_discrete,
-        use_dirichlet=args.use_dirichlet,
-        zero_inflation=args.zero_inflation,
-        gate_prior=args.gate_prior,
-        delta=args.delta,
+        use_zeroinflate=args.zeroinflate,
         loss_func=args.likelihood,
-        dirichlet_mass=args.dirichlet_mass,
         nn_dropout=args.layer_dropout_rate,
         post_layer_fct=args.post_layer_function,
         post_act_fct=args.post_activation_function,
         codebook_size=args.codebook_size,
-        latent_dist = latent_dist,
+        z_dist = args.z_dist,
         dtype=dtype,
     )
 

{sure_tools-2.1.4 → sure_tools-2.1.6}/SURE/SURE.py

@@ -1004,24 +1004,18 @@ def parse_args():
         help="the file for the record of cell-level factors",
     )
     parser.add_argument(
-        "-delta",
-        "--delta",
-        default=0.0,
-        type=float,
-        help="penalty weight for zero inflation loss",
-    )
-    parser.add_argument(
-        "-64",
-        "--float64",
-        action="store_true",
-        help="use double float precision",
+        "-bs",
+        "--batch-size",
+        default=1000,
+        type=int,
+        help="number of cells to be considered in a batch",
     )
     parser.add_argument(
-        "--z-dist",
-        default='normal',
-        type=str,
-        choices=['normal','laplacian','studentt','cauchy'],
-        help="distribution model for latent representation",
+        "-lr",
+        "--learning-rate",
+        default=0.0001,
+        type=float,
+        help="learning rate for Adam optimizer",
     )
     parser.add_argument(
         "-cs",
@@ -1030,6 +1024,13 @@ def parse_args():
         type=int,
         help="size of vector quantization codebook",
     )
+    parser.add_argument(
+        "--z-dist",
+        default='gumbel',
+        type=str,
+        choices=['normal','laplacian','studentt','cauchy','gumbel'],
+        help="distribution model for latent representation",
+    )
     parser.add_argument(
         "-zd",
         "--z-dim",
@@ -1037,6 +1038,27 @@ def parse_args():
         type=int,
         help="size of the tensor representing the latent variable z variable",
     )
+    parser.add_argument(
+        "-likeli",
+        "--likelihood",
+        default='negbinomial',
+        type=str,
+        choices=['negbinomial', 'multinomial', 'poisson', 'bernoulli'],
+        help="specify the distribution likelihood function",
+    )
+    parser.add_argument(
+        "-zi",
+        "--zeroinflate",
+        action="store_true",
+        help="use zeroinflation",
+    )
+    parser.add_argument(
+        "-id",
+        "--inverse-dispersion",
+        default=10.0,
+        type=float,
+        help="inverse dispersion prior for negative binomial",
+    )
     parser.add_argument(
         "-hl",
         "--hidden-layers",
@@ -1070,20 +1092,6 @@ def parse_args():
         type=str,
         help="post functions for activation layers, could be none or dropout, default is 'none'",
     )
-    parser.add_argument(
-        "-id",
-        "--inverse-dispersion",
-        default=10.0,
-        type=float,
-        help="inverse dispersion prior for negative binomial",
-    )
-    parser.add_argument(
-        "-lr",
-        "--learning-rate",
-        default=0.0001,
-        type=float,
-        help="learning rate for Adam optimizer",
-    )
     parser.add_argument(
         "-dr",
         "--decay-rate",
@@ -1105,47 +1113,10 @@ def parse_args():
         help="beta-1 parameter for Adam optimizer",
     )
     parser.add_argument(
-        "-bs",
-        "--batch-size",
-        default=1000,
-        type=int,
-        help="number of cells to be considered in a batch",
-    )
-    parser.add_argument(
-        "-gp",
-        "--gate-prior",
-        default=0.6,
-        type=float,
-        help="gate prior for zero-inflated model",
-    )
-    parser.add_argument(
-        "-likeli",
-        "--likelihood",
-        default='negbinomial',
-        type=str,
-        choices=['negbinomial', 'multinomial', 'poisson', 'gaussian','lognormal'],
-        help="specify the distribution likelihood function",
-    )
-    parser.add_argument(
-        "-dirichlet",
-        "--use-dirichlet",
+        "-64",
+        "--float64",
         action="store_true",
-        help="use Dirichlet distribution over gene frequency",
-    )
-    parser.add_argument(
-        "-mass",
-        "--dirichlet-mass",
-        default=1,
-        type=float,
-        help="mass param for dirichlet model",
-    )
-    parser.add_argument(
-        "-zi",
-        "--zero-inflation",
-        default='exact',
-        type=str,
-        choices=['none','exact','inexact'],
-        help="use zero-inflated estimation",
+        help="use double float precision",
     )
     parser.add_argument(
         "--seed",
@@ -1187,29 +1158,23 @@ def main():
     input_size = xs.shape[1]
     cell_factor_size = 0 if us is None else us.shape[1]
 
-    latent_dist = args.z_dist
-
     ###########################################
     sure = SURE(
         input_size=input_size,
         cell_factor_size=cell_factor_size,
         inverse_dispersion=args.inverse_dispersion,
-        latent_dim=args.latent_dim,
+        z_dim=args.z_dim,
         hidden_layers=args.hidden_layers,
         hidden_layer_activation=args.hidden_layer_activation,
         use_cuda=args.cuda,
         config_enum=args.enum_discrete,
-        use_dirichlet=args.use_dirichlet,
-        zero_inflation=args.zero_inflation,
-        gate_prior=args.gate_prior,
-        delta=args.delta,
+        use_zeroinflate=args.zeroinflate,
         loss_func=args.likelihood,
-        dirichlet_mass=args.dirichlet_mass,
         nn_dropout=args.layer_dropout_rate,
         post_layer_fct=args.post_layer_function,
         post_act_fct=args.post_activation_function,
         codebook_size=args.codebook_size,
-        latent_dist = latent_dist,
+        z_dist = args.z_dist,
         dtype=dtype,
     )
 

{sure_tools-2.1.4 → sure_tools-2.1.6}/SURE_tools.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: SURE-tools
-Version: 2.1.4
+Version: 2.1.6
 Summary: Succinct Representation of Single Cells
 Home-page: https://github.com/ZengFLab/SURE
 Author: Feng Zeng

{sure_tools-2.1.4 → sure_tools-2.1.6}/setup.py

@@ -5,7 +5,7 @@ with open("README.md", "r") as fh:
 
 setup(
     name='SURE-tools',
-    version='2.1.4',
+    version='2.1.6',
     description='Succinct Representation of Single Cells',
     long_description=long_description,
     long_description_content_type="text/markdown",